Decreasing sensitivity of human populations to temperature variability during 50–10 ka in China

Abstract

Understanding how ancient human population size responded to climate change is crucial for studying environmental impacts on human survival and migration. Summed probability distributions (SPD) of archaeological radiocarbon dates have been widely used as proxies for population levels, though previous research for China has mainly focused on the Holocene. The SPD method, however, has known limitations and was challenged by the recently developed Radiocarbon-dated event count (REC) method. Here we investigate the spatiotemporal relationship between climate change and population dynamics in China from 50 to 10 ka using both REC and SPD methods, alongside kernel density analysis for spatial patterns. We find significant correlations between climate variability and population changes through time, which are not detectable with the traditional SPD approach. Notably, we observe a decreasing sensitivity of population size to temperature over time, particularly during 40-35 ka, suggesting growing human resilience and adaptation to environmental stressors, possibly a result of technological advancements. The spatial analysis reveals positive effect of temperature and negative effect of precipitation seasonality on population density, while NPP does not emerge as a major predictor. Population distribution shifted from a northern concentration before the Last Glacial Maximum (LGM) to a more balanced distribution between northern and southern China during and after the LGM, suggesting that migration to regions with milder climate fluctuations may also contribute to the decreased temperature sensitivity through time. These findings highlight the ability of improved statistical methods with temporal and spatial perspectives to capture complex human-environment interactions, providing insights into human adaptability during prehistory.